Self-driving vehicles and the trolley problem

Recent news more than answers the skeptics who said it would take many years until the technology and legislation were ready: Budweiser has just made its first shipment of 50,000 beer cans in a fully self-driving Volvo truck at autonomous driving level 4. A startup created by ex-Googlers and newly acquired by Uber has just sent its first delivery out on the road between two American cities, with a driver sitting behind the wheel who might just as well not have been there.

After a certain amount of intrigue, Tesla has announced that from now on, all its vehicles will include all the necessary kit to allow them to be fully autonomous, again, at Level 4: $8,000 worth of sensors and equipment ready to activated by software that continues to evolve, but when ready, will be installed by a simple upgrade without the need to take the car to a dealer (Tesla, incidentally, has its own dealerships, changing among other things, existing distribution models).

If you’re still skeptical, have a look at this video in which a Tesla drives itself out of its owner’s garage, picks him up, takes him to the shops, leaves him at the door and then goes off in search of a parking spot:

Meanwhile, manufacturers such as BMW, Volvo, Daimler and many others continue to progress toward the new automotive frontier, while a French company completes a round of funding to continue developing self-driving buses, and doubts about Apple’s project are countered by proof that has teams working on developing a self-driving operating system.

And in a new twist to the supposed ethical dilemma, Mercedes has announced that in case of accident, its self-driving vehicles will give priority to the protection of the driver, not bystanders. Along side Volvo’s declaration that it will accept full liability in the event of accidents involving its autonomous vehicles, this is one of the most radical statements we’ve seen in this new era of the automotive industry.

Actually, I’m completely fed up with the tiresome “trolley problem” and skeptics who claim that it will delay the arrival of autonomous vehicles until God knows when. The “trolley problem” is a textbook example of a “non-problem”, whereby we delay the introduction of a technology that can save millions of lives simply because we do not know what to do about something that is never going to happen.

The possibility that a vehicle would ever be in a situation where there is no other choice but to kill a pedestrian or kill the driver is so unlikely as to be not worth thinking about. It might be a fascinating dilemma for ethicists, but in real life, when vehicles are bristling with high quality sensors able to understand what is going on around them, we would do better to worry about other things.

In the event of an emergency today, drivers respond depending on the myriad factors in play at the time. The question we should be thinking about is what happens when such situations no longer happen because the vehicle always knows what is going on at all times? In short, there is no Trolley Problem: no vehicle would ever find itself in such a radically binary situation. If that’s the best the skeptics can come up with, the best thing to do is ignore, learn from the past, continue testing and, soon, get on with marketing self-driving vehicles.

Next question…


(En español, aquí)